A problem that arises with the use of serial block oriented data processing techniques is in the interpretation of the serial data stream. The interpretation is used to determine the control signals that are required to process the serial data stream on a block by block basis.
For example, image processing algorithms typically operate on discrete portions of an image. Each discrete portion is referred to as a block of image data. However, in that multiple blocks can simultaneously exist in a processing pipeline, the supervision of pipeline control can represent a costly overhead factor to a host data processing system or to a dedicated co-processor.
Image compression and decompression (CODEC) techniques, such as those referred to as the Joint Photographic Experts Group (JPEG) and the Motion Picture Experts Group (MPEG), make use of a discrete cosine transform (DCT) function. These techniques divide an image into many small areas, referred to as blocks. The blocks are then processed sequentially within a process-pipeline that includes high speed hardware for implementing the DCT-related functions.
During the processing of such an image, it is often desirable to dynamically modify processing control parameters for the blocks.
It is thus an object of this invention to provide a technique for modifying block control parameters in a dynamic manner.
It is a further object of this invention to provide a technique that provides automatic flexible pipeline control on a per-block basis through the use and interpretation of the contents of a pipeline data buffer.